Research team and available data. This application represents a multi-center collaborative effort using data and specimens from on-going population-based studies to identify and replicate genetic loci that affect the incidence of myocardial infarction (Ml), stroke, and heart failure (HF). The setting is the Cardiovascular Health Study (CHS), an NHLBI-funded cohort study of risk factors for heart disease and stroke among older adults. Replication efforts for Ml and stroke use the Heart and Vascular Health (HVH) case-control study among members of Group Health Cooperative. Data and biologic specimens, from both CHS and HVH, are available on about 1690 MIs, 1110 strokes, and 870 HF events. The research team includes experts in epidemiology, cardiovascular disease, genetics and biostatistics from the University of Washington, the University of Vermont, Cedars-Sinai Institute of Medical Genetics, and the CHS Steering Committee. Context and aim. Linkage studies and candidate-gene approaches have had limited success in identifying common patterns of genetic variation that influence the risk of cardiovascular events. Advances in technology have made it possible to conduct whole-genome (WG) association studies in unrelated populations. The primary aim (aim 1) is to identify 10 underlying genetic variants associated with the risk of each of three major cardiovascular events, Ml, stroke and HF (total of 30 variants). The secondary aim (aim 2) is to configure the design so that the WG scans are conducted on a large random sample of CHS participants. These WG scans can then be used to generate lists of risk markers for the large number of other high-quality phenotypes available in CHS. The secondary WG analyses can then be used as the basis for future genetic studies either within CHS or by other investigators in other populations. Methods. The proposed study has 3 major parts, a two-stage design in CHS plus an external replication study for Ml and stroke. Among the 5888 CHS participants, 4056 were free of all clinical cardiovascular disease at baseline. In the first stage, a random sample of 2000 participants will be selected for WG scans with the Illumina HumanHap300 BeadChip. Cohort analyses will identify the 400 most interesting regions for each of the three major outcomes, Ml, stroke and HF (total of 1200 regions). In the second stage, HapMap data will be used to select an additional 4 SNPs for each of the 1200 regions, and these SNPs plus the originally identified high-signal SNP (6000 across all 3 outcomes) will be genotyped in the other 2000 CHS participants. Cohort analyses will identify the 12 to 20 most interesting regions for replication. In the third part, the best regions identified for Ml and stroke will be replicated in the large HVH case-control study. The proposed three-part study is efficient, has excellent power to detect small to modest-sized hazard ratios, provides a large sample of WG scans for aim 2, and includes an external replication. (End of Abstract)